[GHC] #2893: Implement "Quantified contexts" proposal

[GHC]

#2893: Implement "Quantified contexts" proposal
---------------------------------+------------------------------------------
    Reporter:  porges            |       Owner:                  
        Type:  feature request   |      Status:  new             
    Priority:  normal            |   Milestone:  _|_             
   Component:  Compiler          |     Version:  6.10.1          
    Keywords:  proposal          |          Os:  Unknown/Multiple
Architecture:  Unknown/Multiple  |     Failure:  None/Unknown    
  Difficulty:  Unknown           |    Testcase:                  
   Blockedby:                    |    Blocking:                  
     Related:                    |  
---------------------------------+------------------------------------------
Changes (by sjoerd_visscher):

 * cc: sjoerd@… (added)


Comment:

 I don't see why this needs "local instances". The instances are not the
 problem, we can already do polymorphic instances, i.e. instance Monoid
 [a]. Even specifying the forall constraint now (accidentally) works in GHC
 7.4.1.

 As I see it, the problem is with matching the need for a constraint with
 the available constraints or instances. See #7019, the error is "Could not
 deduce (Monoid [b]) from the context (forall a. Monoid [a])".

 Even g from the previous comment now compiles (with ConstraintKinds and
 UndecidableInstances), except when evaluating it, you get the error:
 "Could not deduce (forall b. Foo [b]) arising from a use of `g'".

 So it seems that with implementing ConstraintKinds, there has been made a
 big step towards an implementation of quantified contexts.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2893#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler